Mavic 3 Pro 60 FPS v 24

[trisen1981]

Hello - I was wondering if there is image quality difference between 24 FPS and 60 FPS on Mavic 3 Pro.

Technically in 24 FPS less data should be written on the card - and potentially less compression -> increased BitRate per second

But is it actually happens or image quality - mostly noise is same in both 24 and 60 FPS at 4k

Has anyone done some comparison?

 

[GadgetGuy]

The bitrate determines the amount of data per frame in the video. If you intend to extract still frames from your video, you'll want 24 fps, not 60 fps. 60fps is intended for slow motion. However, you'll actually have the best results with 30 fps because it most closely matches a multiple of the Hertz refresh rate on mobile devices used for consumption of video.

60fps also loses a stop of light in low light, as the slowest shutter speed becomes 1/60 rather than 1/30. Don't use ND filters for a cinematic effect if you intend to extract usable still frames, as the desired motion blur ruins still frames.

 

[jephoto]

The bitrate determines the amount of data per frame in the video. If you intend to extract still frames from your video, you'll want 24 fps, not 60 fps. 60fps is intended for slow motion. However, you'll actually have the best results with 30 fps because it most closely matches a multiple of the Hertz refresh rate on mobile devices used for consumption of video.

60fps also loses a stop of light in low light, as the slowest shutter speed becomes 1/60 rather than 1/30. Don't use ND filters for a cinematic effect if you intend to extract usable still frames, as the desired motion blur ruins still frames.

Glad you explained this. In the past when I knew I might want to extract still frames I assumed that 60 FPS was the way to go as I’d have a 1/120 shutter speed for sharper stills. Not realizing there would be any drop in quality on the still image.

Or am I reading that wrong and you’re just referring to not being able to go slower than 60th of the second if light is low?

 

[trisen1981]

I am not looking to extract still frames but rather want better quality (less noise) for the video.

If you want to extract stills - just not used ND filters at all and go for faster shutter speed - all will be crystal clear and sharp without any motion blur at all

Tody I recorded some test with 60 FPS and 30 FPS

30 FPs bitrate 112565 kbps
60 FPS 135494 kbps

So double the frame rate but the bitrate is not doubled - that makes me thing that 30 FPS supposed to be a bit better quality that 60 FPS - however visually I hardly see much difference

 

[akdrone]

Using 24 frames per second opens you up to significant issues with judder. Do a web search on that term and you will get an appreciation as to why shooting 30 or 60 frames per second is almost invariably a better choice than 24. 24 was a standard when there was film, but in today’s age when we view videos on monitors in a 60 Hz country 24 simply doesn’t work as well.

 

[jephoto]

Good info. Thx.

 

[Kilrah]

Bitrate i.e. amount of data per second is the same, there are more changes to track at 60fps but each of these changes is smaller so it should pretty much even out.

 

[Droning on and on...]

I am not looking to extract still frames but rather want better quality (less noise) for the video.

If you want to extract stills - just not used ND filters at all and go for faster shutter speed - all will be crystal clear and sharp without any motion blur at all

Tody I recorded some test with 60 FPS and 30 FPS

30 FPs bitrate 112565 kbps
60 FPS 135494 kbps

So double the frame rate but the bitrate is not doubled - that makes me thing that 30 FPS supposed to be a bit better quality that 60 FPS - however visually I hardly see much difference

Gang, video compression is a rather complex system where bitrate isn't going to scale with framerate, while quality can be just as good.

Data "quantity", and therefore bitrate, depends greatly on the difference between frames. Simplified, a complete frame is captured and compressed using similar compression methods to a jpg (an I-frame), and then for a number of frames the difference between the current frame and the previous frame is calculated, and that data is compressed. When there is very little change frame to frame, the difference "image" is highly compressible (think of it as a lot of runs of zeroed pixels, again just an analogy).

Periodically another I-frame is sent, starting the "baseline" for difference calculations over again. This is a configurable part of the image processing pipeline. I-frames are data intensive, so you want to send them as infrequently as necessary, but then has subsequent frames depart from the last I-frame data size per frame increases as well. It's a balancing act.

There are more complexities, like B-frames that are backward referencing, and other technical tricks, but this makes the salient point to understand why 60fps can be compressed with the same quality as 30fps without increasing the bitrate much... Adjacent frames have even less difference from each other at higher frame rates. Two frames that are nearly identical can result in the latter frame compressing to almost nothing without compromising quality -- think of a symbol that says repeat previous frame (not how it actually is encoded, but makes the point).

The same principal applies to 24 vs. 60 fps. Yes, we'd expect a higher bitrate, but not by much.

 

[Droning on and on...]

Using 24 frames per second opens you up to significant issues with judder. Do a web search on that term and you will get an appreciation as to why shooting 30 or 60 frames per second is almost invariably a better choice than 24. 24 was a standard when there was film, but in today’s age when we view videos on monitors in a 60 Hz country 24 simply doesn’t work as well.

And then there's 50/100fps... **** Europeans, always gotta be different